KR100834436B1 - Method and apparatus for load sharing in wireless access networks based on dynamic transmission power adjustment of access points - Google Patents

Abstract

A system and method for load balancing for a packet based wireless cellular network, wherein the wireless access points are dynamically reconfigured by changing the transmit power level to change the coverage area. This change function is based on localized traffic load or congestion at each individual access point, or based on the collective load experienced by groups of access points.

Description

TECHNICAL AND APPARATUS FOR LOAD SHARING IN WIRELESS ACCESS NETWORKS BASED ON DYNAMIC TRANSMISSION POWER ADJUSTMENT OF ACCESS POINTS}

TECHNICAL FIELD The present invention relates to the connection and wireless networking of a computer network, and more particularly, to a cellular wireless access infrastructure that uses a set of access point devices as gateway nodes for connecting individual client wireless devices to a wired backbone network. It is about.

Radio access networks are often designed cellular and have a plurality of access points (also called base stations) that provide a plurality of access points to a backbone (eg, wired) networking infrastructure. As long as individual wireless client computing devices are located within the communication range of the associated access point (s), individual wireless client computing devices may connect to the network by connecting to one of these access points (or, in theory, the minimum number, or more). Can be. Access points (APs) are designed cellular and have each access point area of coverage defined by an area known as cell size. Thus, the composite structure consists of a plurality of cells. Because the cells overlap with each other, the wireless client device located within the overlap area has a plurality of candidate APs, and thus is associated with one or more access points. On the other hand, if the cells do not overlap, the wireless device has only one candidate access point. Thus, this indicates that the set of candidate access points for the wireless device is a function of the coverage area of the individual APs and can only be changed by varying the coverage areas of the individual APs.

The coverage area of the cell is largely dependent on the transmit power used by the AP. It is a well-accepted law of wireless communication that the larger the transmit power, the larger the communication range, and hence the coverage area of the associated cell (also called "communication range"). Communication interruptions and collisions between adjacent cells are avoided through a variety of techniques, such as frequency division (where adjacent cells communicate in non-overlapping frequency bands), code (where adjacent cells use orthogonal codes to avoid mutual interference). Segmentation, or time division (adjacent cells communicate in non-overlapping time intervals).

Each cell (or equivalently AP) has a predetermined traffic capacity defined by various parameters such as available bandwidth or number of channels. If a large number of wireless devices connect to a single AP, or if some of the connected wireless devices generate excessively large amounts of traffic, the access point suffers from congestion overload, resulting in denial of service or poor service performance. . Since the physical movement pattern of wireless devices or the density of those devices in various physical areas cannot be predicted in advance, access points must have a mechanism for handling sudden increases or decreases in load. These mechanisms are called load-balancing and typically take one of two approaches.

Capacity adaptation: The overload access point may attempt to increase its capacity by borrowing capacity from neighboring APs.

Load adaptation: The overload access point may attempt to reduce its load by forcing some or all of the wireless devices involved to switch to alternative neighboring APs.

The most recent load-balancing in cellular networks uses a capacity adaptation method. Overload access points borrow extra capacity from neighboring APs that are not sufficiently utilized. Examples of adaptive capacity adjustment methods include S.Tekinay and B. Jabbari's "Handover and Channel Assignment in Mobile Cellular Networks" (IEEE Communications Magazine, November 1991) and S. Das, S Sen and R. Jayaram Dynamic load Balancing Strategy for Channel Assignment using Selective Borrowing in Cellular Mobile Environments "(Wireless Networks (3), 1997) and" Increasing Channel Occupancy in Large Scale Mobile Radio Systems: Dynamic Channel Assignment "by D.Cox and D.Reudnick. (IEEE Bulletin on Vehicle Technology, 1973). In these techniques, each access point can communicate on multiple channels at the same time. If overloaded, the access point borrows idle communication channels (essentially additional capacity) from neighboring cells. However, the coverage area of each access point remains unchanged. The methods only work if the individual AP has expensive hardware and specialized software that supports multiple channels at the same time. Many AP implementations (eg, implementations for IEEE 802.11 based wireless LANs) can support only one channel, in which case the capacity of each AP is fixed. However, implementations have flexibility in setting the transmit power level. In this environment, channel borrowing schemes cannot be used. However, using a power control approach, a load-balancing scheme is available.

For example, SVHanly's "An Algorithm for combined Cell-Site Selection and Power Control to Maximize Cellular Spread Spectrum Capacity" (IEEE Journal of Selected Regions for Communications, September 1995) and J.Qiu and J.Mark's "A Other publications, such as "Dynamic Load Sharing Algorithm through Power Control in Cellular CDMA", have discussed the use of power control by an AP while communicating with a set of connected wireless devices. The focus of the system described in these references is to reduce the communication power between the AP and a collection of already connected wireless devices to the minimum level required to maintain communication. This power reduction not only conserves energy, but also reduces interference with neighboring cells. In this communication, it is true that power reduction involves a corresponding reduction in the communication range of the access point.

It is understood in this prior art that the reduction is only performed in communication with the connected wireless device. That is, power control is not performed to control the set of devices that can be legally connected to this access point.

Typically, wireless devices use an intermittent beacon signal to determine the set of possible APs. S.Tekinay and B. Jabbari's "Handover and Channel Assignment in Mobile Cellular Networks" (IEEE Communications Magazine, November 1991) and S.Das, S Sen and R. Jayayaram's "Dynamic load Balancing Strategy for Channel Assignment" The approaches described in the references of using Selective Borrowing in Cellular Mobile Environments "(Wireless Networks (3), 1997) do not mention reducing or increasing the power level of this beacon signal, instead they are accessed by nodes. After connecting to the point, adjust the power level based on the individual packets. Accordingly, the system described in this prior art reference does not solve the traffic load balancing problem.

Thus, there is a need for a new traffic-balancing method for APs in cellular and wireless communication networks.

Until recently, no prior art has been discussed regarding increasing the set of wireless devices capable of connecting to and dynamically increasing the power level of an access point without basis. In other words, the AP transmit power is not treated as a parameter of the load-balancing method.

Therefore, it is highly desirable to provide a load-balancing approach that implements the concept of increasing or decreasing the transmit power in advance by the AP.

The main aspect of the present invention is an indirect and implicit technique for performing load balancing in packet-based wireless cellular networks, to implement dynamic changes in transmit power levels used by individual APs. If the power level of the AP is increased, its coverage area is also increased, which implies that the wireless communication device located outside the communication range is now within the extended cell area.

In addition to the above aspects of the present invention, a load balancing system and method for a packet based wireless cellular network is provided. Indeed, in accordance with the present invention, wireless access points are dynamically reconfigured (eg, by changing the transmit power level) to change the coverage area (also known as "cell communication range"). The change function is based on localized traffic load or congestion at each individual access point, or based on the collective load experienced by groups of access points. This variation in cellular deployment is used as an implicit load-balancing technique. If an overload access point reduces its coverage area, some wireless client devices that were previously in its coverage area no longer exist in its coverage area to connect to alternative access points. Similarly, if an underload access point increases its coverage area, this is an opportunity for wireless devices that were previously outside the communication range to now be in the extended coverage area to switch their connection to this access point. To provide. By changing the access point to which the wireless device connects, the load level on the individual access points can be changed indirectly. Advantageously, this change occurs without the need for additional signaling or messages defined between the access point and the individual wireless devices. The default standard functionality of wireless client devices is sufficient to ensure the necessary change in their connection point.

In the event of overloading a particular cell (equivalently an AP), the present invention proposes to increase (in a controlled manner) the power level of adjacent access points so that they increase the overlap area with the particular cell. As a result of this overlap, a very large number of computer devices find it possible to connect to a wired network through alternative APs, distributing part of the load from the current access point to neighboring APs. Alternatively, in the event of overloading a particular cell, the present invention proposes a technique in which the associated AP reduces its transmit power to reduce the communication range of the associated cell. Thus, nodes located outside of the reduced coverage area (but within the boundaries of the original cell coverage) are now forced to join the alternative AP, thereby loading traffic from the (current) overload access point to the appropriate alternative APs. Will alleviate part of it. Advantageously, the power-adaptation techniques implemented do not require special antenna capabilities and work in relatively inexpensive and inexpensive wireless network installations.

The key of the present invention is to use the dynamic change of the topology (topology) of the cellular network through the dynamic change of power level, and the dynamic change of topology as a response mechanism of load-balancing for variable traffic load. As a result of the dynamic change in cellular network deployment (topology) through dynamic changes in power levels, the network-initiated load-balancing does not require additional signaling messages between the wireless devices and the APs. It is advantageous. In fact, power control techniques can be used both with and without additional signaling messages between wireless client devices and APs. The main advantage of the power regulation method is that it does not require any additional signaling messages (others not enforced by the default functionality of the wireless devices) to perform load balancing. Once the coverage area of different APs is changed, the wireless devices will eventually reconnect to alternative APs.

That is, in the exemplary embodiment, when new wireless devices increase power and attempt to connect to that wireless network, they identify the changed coverage area set and connect to the preferred AP based on the modified coverage area set. will be. The network's attempt to balance traffic loads is completely open to wireless devices.

According to another aspect of the present invention, there is provided a centralized entity (called "Load Balancing Manager") that can be implemented in software to adjust the power level (and thus the communication range) of the AP aggregation. By monitoring the load level of each of the individual APs, the LBM can determine the number of connected users that exceed a threshold, or the amount of traffic that exceeds a certain rate, or an acceptable boundary when a particular access point is loaded. Can be expressed as various conditions such as exceeding processing latency. The LBM instructs adjacent light load APs to increase the transmit power level to increase the area of congestion with the current AP's communication range. As a result of this overlap, one or more wireless devices connected to the current AP will be within communication range of the alternative APs. By switching these devices to these APs, the LBM can reduce the traffic load of the current AP to a more acceptable level.

Another aspect of the invention is that the invention further enables independent power coordination by individual APs with minimal cooperation with neighboring APs and no centralized cooperation. An access point that experiences significant overload simply reduces its cell size and requests neighboring APs to increase their coverage area to relieve the load. Each AP uses a wired network to do this in a fully asynchronous manner with a limited amount of localized signaling (to neighboring APs).

In another aspect of the load balancing technique of the present invention, the process of adjusting the transmit power level (cooperative or independent) is initiated by APs or a wireless computer device. The present invention utilizes both a wireless device initiation trigger (such as an indication of poor service or high latency), or an AP initiation trigger (such as an excessive number of users or an excessive traffic rate) at the start of the residual coordination process. In addition, the load balancing technique uses explicit specific signaling messages that instruct the wireless device to recombine with a particular alternative (or any alternative) AP, or leaves it as a later independent event to facilitate the reconnection process to the alternative AP. You can start

Another aspect of the invention relates to combining power regulation techniques in an AP with additional power regulation techniques in one or more wireless devices. In this method, an overloaded AP may relieve some of the load by simply requiring some of the wireless client devices connected to it to increase its transmission range to connect to neighboring APs. To ensure successful bidirectional communication, this method may require one or more neighboring APs to increase their transmit power.

The final aspect of the invention relates to a method for allowing an LBM or an individual AP to determine the extent to which the power level of an overloaded AP needs to be increased or the power level of an underload AP is required to decrease. In practice, all transmit power level values are associated with a well defined transmission or communication range. In addition, in one embodiment, it is assumed that the location of individual APs is known to the system. Accordingly, the LBM or individual AP determines the distance between neighboring APs. To ensure that no "blind spots" occur due to non-overlapping coverage areas of neighboring APs, the system requires that the sum of the transmission ranges of the two APs is the physical distance between them when the AP's power level is reduced. Ensure to increase the power level of the neighboring AP to the minimum level necessary to ensure that it is equal to or greater than the physical distance between them. In addition, the present invention uses a gradual increase or decrease method so that the power level of the overload AP is gradually increased, and the power level of the light load AP is gradually reduced until the loads are satisfactorily redistributed. Of course, alternative embodiments of the present invention may use different algorithms where more or more increases or decreases occur (e.g., in a overloaded server, the amount of reduced transmit power is proportional to the load of that server). have.

The features, aspects and advantages of the apparatus and method of the present invention will be better understood from the following specification, the appended claims and the following drawings.

1 is a general schematic diagram of a cellular access network, which illustrates the role of APs and shows how various wireless devices connect and communicate with particular APs.

2 illustrates an overload situation occurring at an access point in a cellular environment.

3 shows a solution according to the present invention for adjusting the AP coverage area to perform load balancing and to mitigate traffic imbalance.

4 is a flowchart illustrating the basic sequence of steps performed in the load-balancing method according to the present invention.

FIG. 5 is a flow diagram illustrating the basic sequence of steps performed by a centralized software entity called a load balancing manager, acting as a central control point for the power regulation process, in accordance with an embodiment of the present invention.

6 shows an alternative embodiment of power-based load balancing, where each AP operates autonomously and power control is performed in a distributed fashion.

7 is a flowchart illustrating how actual switching of a wireless client device for an alternative access point is affected.

8 illustrates a logical deployment of a cellular network that includes a Load Balancing Manager (LBM) software entity.

The present invention is directed to a new load balancing mechanism that is one family of load adaptation schemes. The prior art mechanism for load adaptation trusts clear signaling (exchange of messages) between the AP and the wireless client node, so that in reality the wireless node switches to an alternative AP even if it is within the footprint of the current AP. On the other hand, the load-balancing mechanism according to an embodiment of the present invention does not require such clear signaling as described below, but instead reduces the communication range of the AP to exclude the wireless device from the new communication range. Thereby forcing the target wireless device to switch to another AP.

That is, the present invention provides a method wherein a set of access points providing wireless cellular coverage adjusts transmit power to change the cellular topology in response to excessive traffic load at a particular "hot-spot". For example, an overload access point may reduce its coverage area (cell size) only by reducing its transmit power. As a result of this reduction in the coverage area, some wireless client devices that were previously using this access point are then combined again with alternative access points, reducing the traffic load of this access point. In a similar manner, the present invention allows an access point that is underloaded or underused to increase its coverage area so that some wireless client devices that previously existed outside of its coverage area can be accessed. Provide an opportunity to combine with

In addition, the present invention is directed to a method and apparatus that allows this power adjustment process to induce traffic balancing in a wireless access network without the need to employ new signaling or protocols in wireless client devices. This access network can provide load balancing even if individual access points do not have the ability to dynamically change their capacity by techniques that borrow channels or bandwidth from neighboring access points. It also provides a control system that includes a centralized software entity (here called "load balancing manager") to coordinate the process by which individual access points (APs) increase or decrease their transmission range. This central coordination allows the neighboring APs to increase their coverage area if the overload access point reduces its coverage area so that the coverage is free of any "blind spots (out of the coverage area of all access points). Ensure that wireless client devices continue to access the wired networking infrastructure In an alternative embodiment, the load-balancing process can be performed autonomously by each access point, where the access point is only a measure of its load level. To increase or decrease its coverage area based on another alternative embodiment, in which load-by-groups of adjacent APs without a central coordinator communicate between themselves to determine appropriate coordination of transmission range. Equilibration techniques may be performed.

1 illustrates a typical schematic of a wireless cellular access network comprising a group of access points (also called base stations; eg, 101, 102, 103, 104, 105, 106). Each base station has a transmit power that determines its coverage area. These coverage areas are represented by circles (eg, 107, 108, 109, 110, 111, 112) drawn in dashed lines in FIG. 1. For a particular access point, the entire area within the corresponding circle corresponds to its coverage area, and that access point is a suitable candidate for connection to any wireless device located within that coverage area. Individual radio access devices (eg, 113, 114, 115, 116, 117) typically move within this coverage area and connect to one of the access points. In general, a particular wireless device may be located in an overlapping coverage area between a plurality of access points (eg, wireless device 115 may be located within coverage areas 102 and 103 to select from a plurality of alternative APs). Can be).

If a large number of wireless client devices connect to a single access point, or if the devices connected to the access point generate unacceptably large amounts of traffic, the corresponding access point becomes overloaded and degrades performance. . 2 illustrates an overload situation that occurs at an access point in a cellular environment. In general, overload occurs when many wireless client devices move to and access a coverage area of a particular access point, increasing the traffic load associated with that access point. On the other hand, when not serving a large number of wireless client devices at the same time, neighboring access points are underloaded. Thus, FIG. 2 shows an example situation where access point A 201 with coverage area 203 is connected to it (shown by lines 211, 212, 213, 214, 215, respectively). Access point B 202 with wireless devices (W1 205, W2 206, W3 207, W4 208, W5 209), while having a similar coverage area 204 There is only one device W6 210 connected to it (shown by line 216). By switching some of the wireless client devices (eg, W4 208 and W5 209) to access point B 202, the system does not cause an unacceptable increase in the overhead of access point B. The overload of the access point A 201 is reduced.

3 illustrates a method of resolving the overload situation of FIG. 2 through an appropriate change in transmit power (coverage area) of access points. By extending its coverage area, the access point actually pulls some client devices from the other access point to move some traffic from the adjacent overload access point. Thus, in FIG. 3, access point B 302 increases its coverage area (from the corresponding value in FIG. 2), and access point A 301 reduces its coverage area. The wireless client devices (W1 305, W2 306, W3 307) continue to associate with the access point A 301 (311, 312, 313, respectively), and the wireless client devices (W6 310). ] Maintains association 316 with access point B 302. However, compared to FIG. 2, the wireless client devices W4 308, W5 309 switch their combination (314, 315, respectively) to access point B to mitigate the congestion of access point A. FIG. In a preferred embodiment of the invention, the switching of devices W4 and W5 to access point B is only a change in transmit power of APs A and B, without the need for additional signaling from or to the client devices. Is done through. It will be appreciated that according to different embodiments of the present invention, different algorithms may be used to calculate the amount (quantity) in which the transmit power of APs is increased or decreased. For example, there is a well-known one-to-one correspondence between any transmit power level and its associated transmit range. Thus, possible combinations of power reduction at overload APs and power increase at adjacent underload APs are possible, as long as the sum of the transmission ranges is equal to or exceeds the physical distance between the APs.

4 is a flow diagram illustrating functional steps implemented to perform load balancing based on transmit power. Basically, the method involves detecting an overloaded or lightly loaded AP, calculating the necessary changes in the AP's power level to demonstrate their changes, and one or more client devices switching their connections to another neighboring AP. It includes the step of making it. As a result of this switching, the load of the overloaded AP is reduced and the load of the underload AP is increased.

As shown in FIG. 4, the first step 401 continuously monitors the status of traffic loads and / or packet metrics, the step 402 detects an overload on one or more APs, and in step 403 one or more Detect light load conditions at APs. Each of these conditions is used to cause an adjustment of the transmit power (coverage area) at one or more APs. Candidate embodiments for algorithms to adjust individual transmit power at the AP (among various other alternatives that can be configured by the skilled worker), a) always reduce the transmit power of the overloaded AP to the next lower power level, Then increase the transmit power of neighboring APs to ensure proper overlap of the coverage area, and b) reduce the transmit power of the overloaded AP based on the size of the current load or the individual or collective location of the currently connected wireless client devices. It involves making. In essence, greater overload reduces the larger amount of transmit power. Similarly, when most wireless client devices are close to the AP, they reduce a greater amount of transmit power than when they are far away.

As a final step in load balancing, one or more client devices switch their connection 405 to a new AP, changing the distribution and traffic load of the individual APs. Accordingly, by returning to step 401 again, the monitoring process continues. Overload detection at one or more APs is performed through various standard methods. Detection can also be performed at either one or more APs or one or more wireless client devices. For example, a simple form of detection involves tracking the number of users each access point is connected to or the number of bits / sec passing through it, and informing an overload condition if these values exceed certain boundaries. Similarly, a wireless client device monitors various performance metrics associated with its packets (e.g., packet delay, or packet loss ratio) and detects overload or congestion by informing them of an overload if they exceed a certain boundary. Display. Similarly, access points detect when they are used insufficiently, extending the coverage area and allowing additional wireless devices to connect to them.

The actual calculation of the new coverage area and transmit power may be performed on a centralized computer server operating under software (referred to as load balancing manager) control or distributed over many individual APs. As one possibility, an increase or decrease in transmit power may occur in a set of discrete steps. For example, an overload access point reduces the transmit power from the current level to the next lower value in a finite set of power levels allowed, and after a certain time interval, whether or not this reduction is adequate for ensuring load distribution. To judge. Thus, if the AP needs to reduce the power level further, this is repeated. Finally, switching of client devices to alternative APs is performed in various ways. Switching may be forced by sending a special signaling signal from the current AP to the alternative APs, or switching may be canceled if they exceed the reduced range of the current AP, or may provide better performance or coverage. Switching may occur as a result of an intermittent client-initiated check to see if alternative APs exist. In one embodiment, the client device may compare the signal strength of the beacons transmitted and received by the individual APs and switch to one AP with the largest signal strength.

In accordance with the present invention, a centralized load balancing manager (LBM) belongs to a common access area and is positioned to manage the adjustment of transmit power for a set of APs under the control of a common management entity. Details regarding the operation of the load balancing manager are discussed in detail herein with reference to FIGS. 5 and 8.

5 illustrates the method performed by the LBM as part of the load-balancing function. The LBM determines the appropriate load levels for the various APs (step 501) and compares them with policy-dependent thresholds to determine if power coordination is needed to resolve unacceptable high overload or light load situations. It is determined whether or not (step 502). Typically, these load levels are a measure of various well-known performance metrics (e.g., packet transmission delay, packet loss ratio, packet transmission delay change (jitter), AP processor availability, AP memory availability, number of connected wireless devices, etc.). It is calculated by well known techniques using To illustrate, these sets of metrics are described in "Framework for IP Performance Metrics" (IETF, RFC 2330, May 1998, http: // www. V. Paxson, G.Almes, J.mahdavi, M.Matthis). (available at jetf.org/rfc/rfc2330.txt), which is incorporated by reference as described herein. If the current conditions are good, the LBM skips any adjustment and waits for a time interval before checking the conditions of the access network again (step 503). If the current traffic condition indicates that it is necessary to change the power level of the various APs, the LBM first searches for the geographical placement of that AP (step 504) to determine a set of neighboring APs for that particular AP. Neighboring AP information is available in the form of a simple connectivity graph (with an edge between two neighboring APs), or in the form of the exact geographical location of each AP, or other form of location representation. Based on this location information and load level, the LBM calculates a set of APs for which power is to be adjusted (step 505). This calculation can be performed in a variety of ways. In the simplest calculation, the LBM does not specify a definite value for the amount of reduction, but instructs the overloaded AP to reduce the transmit power. In more complex algorithms, the LBM calculates the exact adjustments needed for the power of each AP, managing to ensure that the adjustment of the coverage area does not produce a "blind spot". Finally, the LBM must instruct the individual access point to make appropriate power level adjustments before repeating again after the specified time interval (step 506).

In another embodiment of the present invention, the adjustment of transmit power is performed by the APs themselves, even without a centralized control entity. In a non-intensive embodiment of the present invention, a set of typical steps is shown in FIG. 6, which represents an alternative embodiment of power-based load balancing, where each AP operates autonomously and power control is performed in a distributed manner. . In this method, each AP queries the neighboring APs whether the neighboring AP can change the power level and whether it should change. This approach uses localized load-balancing to switch wireless client devices from the overloaded AP to one of the adjacent light load APs. Thus, as shown in FIG. 6, each access point must first determine whether it is a location that increases or decreases the transmit power (coverage area). To determine this, each AP first obtains current performance metrics, such as the number of users connected or the traffic rate, and compares it with the thresholds (step 601). As a result of this calculation, the AP can determine whether it is currently overloaded (need to reduce the power level), lightly loaded (can increase the power level if needed by some neighboring AP), or the current transmit power level. It may be determined whether it is an operating range that should be kept (step 602). If the AP is in a light load state, the neighboring APs are queried to determine which of them is overloaded and to send some load to the AP (step 603). After obtaining this information, the access point determines whether it should increase the coverage area (step 604). If this light load access point has an overload neighbor AP, this embodiment is performed by increasing the transmit power of the light load AP and reducing the transmit power of one or more adjacent overload APs (step 605). After completing this process, the AP waits for a certain time interval before returning to step 601 (step 609). If the AP is in a state of maintaining its current power level, it only waits for a certain time interval before returning to step 601. However, if the access point is overloaded, it queries the neighboring APs to find out which of them is lightly loaded and can increase the coverage area (step 606). If an adjacent light load AP is found (step 607), then this embodiment may indicate that the overload AP reduces its transmit power and waits at step 609 and begins the process again at step 601, and at least one adjacent AP. Require them to increase their transmit power (step 608)

Advantageously, the load balancing method according to another aspect of the present invention does not directly include a mechanism by which the wireless client device switches its access point. The client device is free to act on this switching using various alternative techniques. 7 illustrates a set of alternative methods that may be used with a power adjustment algorithm that enables a wireless client device to switch APs in accordance with the present invention.

In the embodiment shown in FIG. 7, the current access point sends a clear reconnection command to the wireless device. This reconnection command merely instructs the client device to specify the identification number of the new access point, or to select another alternative access point. In the latter case, the client device is free to use any algorithm that uses a measure of link quality to determine the desired access point from a suitable alternative access point. If the AP does not send a clear reconnection command, an alternative form of reconnection is caused by the client device currently losing connectivity with the AP. In this case, the client is free again to invoke an algorithm to select an alternative AP. It is important to note that many of these algorithms are available and that the present invention does not force the use of a particular one among others. Thus, clients only need to have the ability to evaluate and select desirable APs among alternative candidate APs. This is a fundamental capability that exists in all practically useful client wireless devices.

The exact order of steps that follow in the method shown in FIG. 7 depends on whether the access point generating the load explicitly instructs the client device to switch to another access point (step 701). If the current AP sends an explicit switching command, this command will explicitly include the identification number of the new AP (step 702). In this case, the client device only needs to set the link and network layer protocols to switch its connection to the new specific AP (step 704). If the switching command does not explicitly specify the new AP (just instructing the client device to switch to some other AP), it determines which of them is the best choice and subsequently connects to the newly selected AP. However, the switching mechanism does not explicitly trust certain commands from the current AP to the wireless device. In addition, switching can be activated by other standard techniques. For example, if the reduced coverage area of the current AP causes the wireless client device to lose connection with it (step 705), the client device autonomously searches for new alternative APs, selects a new candidate AP and Will connect (step 707). In addition, switching may occur even if the reduced coverage area of the current AP does not cause a clear loss of connectivity. Most wireless client devices meet certain protocols that intermittently assess the link quality for the current AP as compared to the link quality for alternative APs. Due to the power level reduction of the current AP and the power level increase of the neighboring APs, the client finds the presence of a better link to the alternative AP during this intermittent evaluation procedure. In this case, the client switches to the alternative AP by performing the necessary reconnection protocol. To prevent a plurality of switching instantaneously occurring when neighboring APs adjust the power level without accurate synchronization, the client device then waits for some time interval before determining the next desired AP.

Thus, the present invention does not specify a new method of connecting to or selecting a preferred AP, but merely changes the coverage area of different APs so that load-balancing is achieved by this connection or selection method.

One of the features of the present invention is that the centralized load balancing manager (LBM) does not need to be located in a particular place, only need to have a network connection to the set of APs it controls. 8 illustrates a logical deployment of a cellular network including a load balancing manager (LBM). According to the present invention, the LBM needs to be able to interface with individual APs to determine their load level and instruct them to adjust their transmit power. As such, the physical placement of the LBM within the network is not critical. As long as there is a network connection between the LBM and the individual APs, it can be located anywhere on the cellular access network or the Internet. In particular, there are two possible deployment alternatives for three access points (A 801, B 802, C 803), and LBM with coverage areas 804 805, 806, respectively. do. In a typical embodiment of a practical access network, these APs will be connected to each other via the local area network (LAN) infrastructure 807. This LAN will in turn be connected to the backbone Internet infrastructure 808. In one embodiment of the invention, the LBM 809 may be located within the LAN itself and be able to communicate with the APs using an access network. In an alternative embodiment, the LBM 810 may be located far from the location where the APs are deployed, and thus need to connect to the APs over the Internet. To connect to the Internet, the remotely deployed LBM 810 will use its own local LAN infrastructure 811.

While the present invention has been described and described in detail in connection with exemplary and pre-formed embodiments, those skilled in the art will appreciate that the foregoing and other changes in form and detail are within the scope and spirit of the invention as defined only by the appended claims. It will be understood that it can be done without.

Claims (29)

A method for mitigating traffic load on one or more wireless access points (APs) in a wireless cellular network, the method comprising: a) detecting one of a traffic overload or underload condition at one or more APs; b) the AP querying one or more neighboring APs to ascertain whether a power level should be changed in the neighboring APs; c) changing the size of the associated coverage area of each AP by changing the transmit power level associated with the one or more neighboring APs according to the detected condition; And d) initiating switching of one or more current wireless devices, from a current connection with a first one of the one or more APs to a connection with an alternative neighboring AP, thereby altering the traffic load of the first and alternative APs; Including steps Each AP operates autonomously to perform AP power control in a distributed manner, without depending on the actual measurement of the transmission characteristics of individual devices or APs. How to reduce traffic load. The method of claim 1, The wireless cellular network is packet based, The detecting step, Monitoring one or more of metrics related to traffic and various performance metrics associated with packets transmitted through the AP; And Generating an overload or light load condition signals at the AP when the overload or light load condition is detected at an AP; Traffic load reduction method comprising the. The method of claim 2, The modifying step c) includes either increasing or decreasing the transmit power level associated with one or more APs, And wherein the transmit power is reduced to a lower power level at an overloaded AP and the transmit power at a neighboring AP is correspondingly increased to ensure overlap of coverage areas. The method of claim 3, The detecting step further includes tracking the number of users connected to each access point and generating an overload condition signal when the number of connected users exceeds a threshold. . The method of claim 3, The performance metric includes one or more of packet latency, packet loss rate, and AP processor availability, The detecting step compares one or more of the packet delay time, packet loss rate, and AP processor availability with each threshold, and wherein the values indicative of the packet delay time, packet loss rate, and AP processor availability exceed a threshold. Generating an overload condition signal at the AP. The method of claim 3, The metric associated with the traffic includes the total traffic rate or the number of users at one or more access points, The detecting step compares the total traffic rate or the number of users at one or more access points with a threshold and if the total traffic rate or the number of users at one or more access points exceeds a threshold. Generating an overload indication signal at the AP. delete The method of claim 1, The switching start step d) is Informing another AP of which the client device has been identified to switch, or enabling the client device to discover a new AP and switch to the new AP. A system for mitigating traffic load on one or more wireless access points (APs) in a wireless cellular network, the system comprising: Means for detecting one of a traffic overload or underload condition at one or more APs and generating one or more overload signals or light load signals at the AP upon detecting the overload or light load condition, respectively; Means for communicating the overload and light load signals to one or more neighboring APs; Means for changing a transmit power level associated with the one or more APs, each AP autonomously operating to perform AP power control in a distributed manner, thereby changing the size of the associated coverage area of each AP; And Switching one or more current wireless devices from a current connection with a first one of the one or more APs to an alternative neighboring AP to change the traffic load of the first and alternative APs, Means for performing without depending on the actual measurement of the transmission characteristics of the APs Traffic load reduction system comprising a. Claim 10 was abandoned upon payment of a setup registration fee. The method of claim 9, The wireless cellular network is packet based, The detection means, Means for monitoring one or more of metrics related to traffic and various performance metrics associated with packets transmitted through the AP; Means for determining the overload or light load condition by comparing metric values related to the traffic and packet performance to a threshold value; And Means for generating overload or light load condition signals at the AP when an overload or light load condition is detected at the AP Traffic load reduction system comprising a. Claim 11 was abandoned upon payment of a setup registration fee. The method of claim 10, The means for changing the transmit power level, To either increase or decrease the transmit power level, And wherein the transmit power is reduced to a lower power level at an overloaded AP, and the transmit power at a neighboring AP is correspondingly increased to ensure overlap of coverage areas. The method of claim 11, The transmit power at an individual AP determines an increase in the power level needed to allow certain client devices connected to alternative APs adjacent to the individual AP to switch to the individual AP, or connected to the individual AP. And varying according to the location of the wireless client devices to determine a reduction in power level required to ensure that particular client devices switch to alternative APs adjacent to the respective AP. delete delete Claim 15 was abandoned upon payment of a registration fee. The method of claim 12, The AP generates signals indicative of another AP identified for the client device to switch, or signals that cause the client device to start searching for a new AP to be switched. delete delete A method for changing transmit power levels at individual wireless access points (APs) in a wireless cellular network, the method comprising: a) detecting one of the traffic overload or light load conditions at one or more APs and generating one or more overload or light load signals at the AP upon detecting the overload or light load condition; b) one or more adjacent APs—each AP is arranged to cooperate with the one or more neighboring APs to change the associated transmit power level according to the detected condition, each AP operating autonomously Communicating AP power control in a distributed manner without depending on the actual measurement of the transmission characteristics of the device or APs, changing the size of the associated coverage area of each AP; And c) initiating switching of one or more current wireless devices from a current connection with a first one of the one or more APs to a connection with an alternative neighboring AP. Including, And the traffic load of the first and alternative APs is changed by the switching. Claim 19 was abandoned upon payment of a registration fee. The method of claim 18, The AP transmit power level at an individual AP determines the power level increase needed to allow certain client devices connected to alternative APs adjacent to the individual AP to switch to the individual AP, or connected to the individual AP. And varying according to the location of wireless client devices that determine a reduction in power level needed to ensure that particular client devices switch to alternative APs adjacent to the respective AP. delete Claim 21 was abandoned upon payment of a registration fee. The method of claim 19, The step of changing the transmit power of the AP, Reducing the transmit power of the overloaded AP or increasing the transmit power of the lightly loaded AP. Claim 22 was abandoned upon payment of a registration fee. The method of claim 21, The wireless cellular network is packet based, Monitoring at each AP one or more of metrics relating to traffic and various performance metrics associated with packets transmitted through the AP; Comparing the threshold values related to the traffic and packet performance with a threshold to determine the overload or light load condition; And Generating an overload condition signal at the AP when the overload and light load conditions are detected at the AP Traffic load balancing method comprising the. Claim 23 was abandoned upon payment of a set-up fee. The method of claim 21, The transmit power at an individual AP determines an increase in the power level needed to allow certain client devices connected to alternative APs adjacent to the individual AP to switch to the individual AP, or connected to the individual AP. And varying according to the location of wireless client devices that determine a reduction in power level needed to ensure that particular client devices switch to alternative APs adjacent to the respective AP. Claim 24 was abandoned when the setup registration fee was paid. The method of claim 21, The step of increasing the AP transmit power further comprises increasing the transmit power of the AP beyond the level required to support communication with the currently connected set of devices. Claim 25 was abandoned upon payment of a registration fee. The method of claim 21, The step of increasing or decreasing the AP transmit power further comprises increasing or decreasing the transmit power of one or more wireless client devices that may or may be connected to the AP. Claim 26 was abandoned upon payment of a registration fee. The method of claim 21, Changing the AP transmit power further comprises causing the coverage area of one or more APs to be entirely contained within the larger coverage area of another access point. delete Claim 28 was abandoned upon payment of a registration fee. The method of claim 21, The transmission power at the individual AP with one or more connected wireless client devices generates signals that the client devices will receive, the signals instruct the client devices to join another alternative access point. Traffic load balancing method. Claim 29 was abandoned upon payment of a set-up fee. The method of claim 28, The signals received by the connected wireless devices may be signals that instruct the wireless client device to switch explicitly to an alternative specific access point, or to the wireless client device to allow the wireless client to find and switch to an alternative access point. Contains commands to command, And the commanded wireless device loses its connection with the access point to which the wireless device is currently connected, and then joins an alternative access point.

Images